Many software projects collapse under the weight of their own complexity. During the collapse, these projects go through familiar stages as they die a painful death.
- First come minor schedule slips
- Then a missed milestone
- Then an adjustment to the master schedule extending the end-date a bit
- Then 80-hour weeks for the programmers
- Then the altered schedule does not work either
- Then management looses touch with reality and insists all is fine
- The end-date comes and goes with no working software.
- And then the finger pointing begins to assess blame.
Why does this happen again and again, to competent organizations that are able to complete other types of projects efficiently? The problem is that many of the world’s large software projects are just too darn complicated.
This article introduces the 1-5-10 Rule for software design to help your projects avoid a similar fate.
Full article here….
The FBI has been trying for twelve years to modernize its case file computer system. The current system was antiquated when it was introduced in 1995 and has been widely criticized as one of the reasons the FBI was not able to “connect the dots” in its terrorism investigations in 2000 and 2001.
The first modernization effort, Virtual Case File, ran from 2000 to 2005, cost $170 million and failed completely.
In March 2006, the FBI started the Sentinel project to try again. The original target completion date was December 2009. The target slipped to June 2010, then September 2010, then September 2011 and then January 2012. Recently, the completion target slipped again to May 2012. The cost has also risen, from $425 million to $451 million, dwarfing VCF.
Sentinel will not succeed, given its current definition, budget and schedule. Even allowing for some missing features, an additional 10% budget increase, and another four-month slip to September 2012, Sentinel will fail. I strongly suspect that in May 2013 it still will not be fully operational with high quality.
Full article here….
There are many troubled software projects in the world right now. There always are. If you ask the programmers and line managers for one of these projects why things are going wrong, they will probably say something like…
- We don’t have enough time and money.
- Upper management is trying to squeeze a year of work into every quarter, and they are making each of us do the work of two people.
- If we just had a more realistic schedule and more help, we could get this thing done.
It certainly is true that some software projects fail because of too little time and money. Startup companies often face the challenge of finishing a commercial-grade program with limited resources.
But lack of time and money is not the reason for the colossal software failures that make headlines. The reason for truly expensive failures is actually the opposite. Big software projects usually go off track because they have too large a budget, too many people working on them and are too important.
Full article here: